Microsoft Fabric is a single analytics platform that bundles together most of the tools a business needs to move, store, model and report on data. If you have ever found yourself with a Power BI workspace over here, an Azure Data Factory project over there, a Synapse warehouse somebody set up in 2021 and a Databricks notebook nobody wants to touch, Fabric is Microsoft's answer to that mess. It puts the lot into one product, with one bill, one storage layer and one place to manage permissions.
That is the short version. The longer version is more interesting, because Fabric is not really a brand-new product. It is a re-packaging of services Microsoft already had, glued together with a shared storage layer called OneLake and a single capacity-based billing model. Whether that re-packaging is a big deal for your business depends on what you are doing today.
What is actually inside Microsoft Fabric?
Fabric is made up of several workloads that used to be sold separately. Data Factory handles ingestion and pipelines — pulling data out of source systems on a schedule. Data Engineering gives you Spark notebooks for heavier transformation work. Data Warehouse is a T-SQL warehouse for people who think in SQL. Real-Time Intelligence handles streaming data. Data Science is the notebook environment for machine learning. And Power BI sits on top of all of it for reporting.
The clever bit is OneLake. Every workload writes to the same storage in the same open format (Delta Parquet), so a table created by a data engineer is immediately queryable from the warehouse, visible to Power BI through DirectLake, and available in a notebook without copying anything. In the old world you would have had three or four copies of the same data sitting in different services. In Fabric, one copy is the rule.
How does Fabric relate to Power BI?
Power BI is now a workload inside Fabric rather than a standalone product. Existing Power BI Pro and Premium licences still work, and nothing breaks if you do nothing. But Fabric adds a new way of feeding Power BI from OneLake called DirectLake, which is genuinely fast and avoids importing data into a dataset at all. For organisations that have been bumping into Power BI dataset size limits, that alone is worth a look.
If you only use Power BI for a handful of finance reports built on Excel exports, Fabric will feel like overkill. If you are running gateway refreshes that take an hour, juggling a dozen datasets that all pull from the same source, or paying for Premium capacity that sits idle most of the day, Fabric starts to look more attractive.
Who is Fabric actually for?
In our experience, three groups get the most out of Fabric in its first year of life. The first is mid-sized businesses (200–2,000 staff) that have outgrown Excel and a single Power BI workspace but cannot justify a dedicated data platform team. Fabric gives them almost everything in one capacity, with no infrastructure to manage.
The second is organisations already deep into Azure who are tired of wiring Data Factory, Synapse and Power BI together by hand. Fabric simplifies the architecture and the billing, even if the underlying engines are familiar.
The third is teams whose data is genuinely big — tens of millions of rows refreshed daily — where Power BI's import mode is starting to creak. DirectLake on Fabric is the answer Microsoft has been promising for years.
Who is it not for? Small businesses running fewer than ten reports on finance data that fits in a spreadsheet. The smallest Fabric capacity (F2) starts at around £210 per month list price in the UK and functionally replaces Power BI Premium Per User at £18 per head. Below roughly twelve report consumers, sticking with Power BI Pro almost always works out cheaper.
What does it cost?
Fabric is sold as capacity units (CUs), priced per hour. The smallest SKU, F2, lists at about $262 per month in the UK if run 24/7, but you can pause it. The next step, F4, doubles that. From there it scales in regular doublings up to F2048. Most mid-sized organisations we work with sit happily on F8 or F16.
The trick most teams miss in their first month: capacity is shared across every workload, so a runaway notebook can starve Power BI reports of compute and vice-versa. Monitoring the Capacity Metrics app from day one is not optional.
The honest trade-offs
Fabric is genuinely promising, but it is still maturing. Three things we tell clients to be aware of before they commit. First, the SQL endpoint on a Lakehouse is read-only — if your team is used to writing back to a warehouse from a stored procedure, you will need the Warehouse workload, not the Lakehouse. Second, fine-grained security (row-level, column-level) is improving every release but is not yet at SQL Server parity. Third, the documentation moves faster than most of us can read it, so build a habit of checking the release notes monthly.
None of those are deal-breakers. They are the normal trade-offs of adopting a platform that is two years old rather than ten. If you want a guided look at whether Fabric fits your setup, our Microsoft Fabric consultancy page covers how we usually approach the first engagement, and we are happy to give a straight answer rather than push you onto a platform you do not need.
Frequently asked questions
Is Microsoft Fabric replacing Power BI?
No. Power BI is now a workload inside Fabric, but existing Power BI Pro and Premium licences continue to work exactly as before. Nothing breaks if you do nothing.
Do I need Microsoft Fabric if I already use Power BI?
Not unless you have outgrown what Power BI can do on its own — large datasets, multiple source systems, long refreshes, or a real need for a shared lakehouse. For small finance reporting, Power BI Pro is still the right answer.
What is the cheapest Microsoft Fabric capacity?
The smallest SKU is F2 at roughly £210 per month list price in the UK if left running. You can pause capacities when not in use, which lowers the effective cost for development and testing.
Want to talk this through with someone?
We are an independent UK Power BI and Microsoft Fabric consultancy. Honest opinions, fair prices, no sales pressure.